Electronic 'crowd behaviour' revealed in semiconductors

Like crowds of people, microscopic particles can act in concert under the right conditions. By exposing crowd behaviour at the atomic scale, scientists discover new states and properties of matter.

Now, ultrafast lasers have revealed a previously unseen type of collective electronic behaviour in semiconductors, which may help in the design of optoelectronic devices. The work has taken place at JILA, a joint venture of the National Institute of Standards and Technology (NIST) and the University of Colorado at Boulder.

Design of optoelectronic devices, like the semiconductor diode lasers used in telecommunications, currently involves a lot of trial and error. A designer trying to use basic theory to calculate the characteristics of a new diode laser will be off by a significant amount because of subtle interactions in the semiconductor that could not be detected until recently.

To shed light on these interactions, the JILA team used a highly sensitive and increasingly popular method of manipulating laser light energy and phase to reveal the collective behaviour of electronic particles that shift the phase of any deflected light. Their work is an adaptation of a technique that was developed years ago by other researchers to probe correlations between spinning nuclei as an indicator of molecular structure.

In the latest JILA experiments, a sample made of thin layers of gallium arsenide was hit with a continuous series of three near-infrared laser pulses lasting just 100 fs each. Trillions of electronic structures called excitons were formed.

By tinkering with the laser tuning and analysing how the semiconductor altered the intensity and phase of the light, the researchers identified a subtle coupling between pairs of excitons with different energy levels, or electron masses.

The experimental data matched advanced theoretical calculations of the electronic properties of semiconductors, confirming the importance of the collective exciton behaviour — and demonstrated the superiority of those calculations over simpler models of semiconductor behaviour.

The work may help researchers better predict optoelectronic device characteristics; not only the magnitude of the emissions signals but also the phase, which is especially significant in optics.